mirror of
https://github.com/c64scene-ar/llvm-6502.git
synced 2024-12-21 16:31:16 +00:00
d6aa02de10
to a few tests where it is required for the expected transformation. git-svn-id: https://llvm.org/svn/llvm-project/llvm/trunk@77290 91177308-0d34-0410-b5e6-96231b3b80d8
34 lines
1.6 KiB
LLVM
34 lines
1.6 KiB
LLVM
; RUN: llvm-as < %s | opt -instcombine | llvm-dis | grep {icmp eq i32 %indvar, 0}
|
|
; PR1978
|
|
target datalayout = "e-p:32:32:32-i1:8:8-i8:8:8-i16:16:16-i32:32:32-i64:32:64-f32:32:32-f64:32:64-v64:64:64-v128:128:128-a0:0:64-f80:128:128"
|
|
target triple = "i686-apple-darwin8"
|
|
%struct.x = type <{ i8 }>
|
|
@.str = internal constant [6 x i8] c"Main!\00" ; <[6 x i8]*> [#uses=1]
|
|
@.str1 = internal constant [12 x i8] c"destroy %p\0A\00" ; <[12 x i8]*> [#uses=1]
|
|
|
|
define i32 @main() nounwind {
|
|
entry:
|
|
%orientations = alloca [1 x [1 x %struct.x]] ; <[1 x [1 x %struct.x]]*> [#uses=2]
|
|
%tmp3 = call i32 @puts( i8* getelementptr ([6 x i8]* @.str, i32 0, i32 0) ) nounwind ; <i32> [#uses=0]
|
|
%tmp45 = getelementptr inbounds [1 x [1 x %struct.x]]* %orientations, i32 1, i32 0, i32 0 ; <%struct.x*> [#uses=1]
|
|
%orientations62 = getelementptr [1 x [1 x %struct.x]]* %orientations, i32 0, i32 0, i32 0 ; <%struct.x*> [#uses=1]
|
|
br label %bb10
|
|
|
|
bb10: ; preds = %bb10, %entry
|
|
%indvar = phi i32 [ 0, %entry ], [ %indvar.next, %bb10 ] ; <i32> [#uses=2]
|
|
%tmp.0.reg2mem.0.rec = mul i32 %indvar, -1 ; <i32> [#uses=1]
|
|
%tmp12.rec = add i32 %tmp.0.reg2mem.0.rec, -1 ; <i32> [#uses=1]
|
|
%tmp12 = getelementptr inbounds %struct.x* %tmp45, i32 %tmp12.rec ; <%struct.x*> [#uses=2]
|
|
%tmp16 = call i32 (i8*, ...)* @printf( i8* getelementptr ([12 x i8]* @.str1, i32 0, i32 0), %struct.x* %tmp12 ) nounwind ; <i32> [#uses=0]
|
|
%tmp84 = icmp eq %struct.x* %tmp12, %orientations62 ; <i1> [#uses=1]
|
|
%indvar.next = add i32 %indvar, 1 ; <i32> [#uses=1]
|
|
br i1 %tmp84, label %bb17, label %bb10
|
|
|
|
bb17: ; preds = %bb10
|
|
ret i32 0
|
|
}
|
|
|
|
declare i32 @puts(i8*)
|
|
|
|
declare i32 @printf(i8*, ...)
|